Abstract
The Navy's EPANALOG (Northeastern Pacific Anlog Tropical Cyclone Tracker) forecast program is introduced. EPANALOG selects analog tropical cyclones from a 25-year northeastern Pacific Ocean history. The selected analog tracks, statistically adjusted for position, vector motion, and date differences between them and the recent history of the tropical cyclone being forecast, are composited into a single forecast track. Verifications of EPANALOG forecasts to 96 h, as initiated from best-track positions, are shown for randomly selected historical cases with a Monte Carlo simulation of initial position inaccuracies, as well as for forecasts generated from 1973 operational cyclone positions. The latter are intercompared with a homogeneous set of objective persistence and MOHATT forecasts as well as subjective OFFICIAL forecasts for the 24, 48, and 72 h intervals. The 1973 EPANALOG accuracy is shown to generally excel that of the existent techniques for all forecast intervals tested.
